Is it possible? 1/2 FSB75 --> 1/3 FSB99?

Batch codes, RAM specs, BIOS settings, etc..
Post Reply
Florek
Posts: 6
Joined: Fri Oct 18, 2002 3:02 pm
Location: Gdansk, Poland

Is it possible? 1/2 FSB75 --> 1/3 FSB99?

Post by Florek »

Hi everyone. I'm quite new here so sorry if this is a repost. I got a little question. Is it possible to change CPU speed from 466(Fsb75)=525Mhz 1/2 to 466(FSB99)=xxx 1/3?? I can't go over FSB78 'couse harddisk crashes :/
Florek
Posts: 6
Joined: Fri Oct 18, 2002 3:02 pm
Location: Gdansk, Poland

Re: Is it possible? 1/2 FSB75 --> 1/3 FSB99?

Post by Florek »

Raffnix wrote:
Florek wrote:Hi everyone. I'm quite new here so sorry if this is a repost. I got a little question. Is it possible to change CPU speed from 466(Fsb75)=525Mhz 1/2 to 466(FSB99)=xxx 1/3?? I can't go over FSB78 'couse harddisk crashes :/
Hi Florek!

If you mean overclocking your Celeron 466 on an Abit BP6, you only have to reach fsb=92 MHz, course the divider (anyone should correct me) changes to 1/3 at this speed.
I could imagine, with 2,1 - 2,3 Volt your Celeron may work with 644.
But i don't believe it.
Why not? :wink: What are the limitations? Temperature?
davd_bob
Confused
Posts: 1043
Joined: Fri Feb 13, 2004 2:30 am
Location: Houston, TX

Post by davd_bob »

Florek,
(and anyone else with the same questions)
The FSB/agp/pci/ide relationships seem complicated at first but once you get used to them its pretty easy on the BP6.

About FSB/divider/cards/HDs ect.
FSB 66-91 are on 1/2 divider. That means everything on the board is overclocked between 67 and 91 FSB. Harddrives often choke above 78FSB(actually the drive designed for 33 is running at 37).
FSB 92 to 123 is 1/3 divider. 92-98 means everyghing on the board is UNDERclocked. Cards and hds that work at 66 is should be stable at 100. That doesn't include the ram or cpu of course because they run off full FSB speed.
FSB 124-133 is 1/4 divider. Don't bother cause only super hero overclockers can get a BP6 to run at those speeds.

About RAM
The faster the FSB the more you gain on the exchange between the CPU and RAM. That alone is one of the single most signifacant places for preformance gains on the average system.
Make sure to use ram rated to the FSB you use. Higher rated is also ok. I use PC-133 on my BP6.
There is a CL rating on ram. Im over simplefyng to say it like this but basicly the ram takes one clock click for each CL increment. CL2 is faster then CL3. If ram is rated for PC-100 that means its CL3 unless marked otherwise. PC-133 CL3 will almost always run at CL2 on 100FSB. ECC ram is more expensive and has extra error correction capibility, but that extra feature costs one CL incirement on the speed. Unless you are using it in a server or for some highly important task dont bother with ECC ram. Before anyone asks, no...DDR ram does not work on a BP6.

About the Celeron
The Mendocino core built on 25micron technology hits a wall between 550 and 625MHz. Pretty much any of them will run 550 nomater what speed they are "rated" for but almost none will go 625 mhz. If your system is running stable at 2.0volts try drop it to 1.85 and you may get a pleasant supprise.

About the BX(with the "Greenie" on it)
The BX440 was designed for 66 or 100FSB. The Softmenu II will allow speeds that go up to 133 but that exceeds by quite a bit the current usage and heat disipation of the BX.
For some reason Abit didn't use HS greese under green HeatSink on the BX chipset. Most of us that overclock our BP6 do 2 things. Use greese on the chipset and add a fan to improve cooling. I personally spent $12 on a chipset cooler for video cards and it works great and looks really cool to.
Good luck.
There are *almost* no bad BP6s. There are mostly bad caps.

No BP6s remaining
Athlon 2800
Sempron 2000
ViaCPU laptop with Vista.(Works great after bumping ram to 2Gig)
P-III 850@100
BCN
Posts: 508
Joined: Tue Nov 26, 2002 9:50 am
Location: Barcelona, Spain
Contact:

Post by BCN »

it might just be, that it does not crash at 78FSB (550Mhz for cel466) because of your HDD - I eally doubt that, it surely crashes because of your CPU, just get it more volts (up to 2.3V should be more than enough for this speed
Dual C366@550MHz 1.90V :) (History)
yet single PIII-S 512Kb L2 cache at 1400MHz@700MHz
BP6 (not modded yet)
256MB PC133 C2
GF4Ti4200-8x
Maxtor 2x60Gb - all on promise ATA133
Lite-On LTR 40125S@48125W!!!
Plus P4 system
davd_bob
Confused
Posts: 1043
Joined: Fri Feb 13, 2004 2:30 am
Location: Houston, TX

Post by davd_bob »

your best bet is scrounge up a 366s then run at 100FSB.
Mine works great at 1.9 volts with dual 366@550.
There are *almost* no bad BP6s. There are mostly bad caps.

No BP6s remaining
Athlon 2800
Sempron 2000
ViaCPU laptop with Vista.(Works great after bumping ram to 2Gig)
P-III 850@100
phaedrus
Posts: 75
Joined: Tue Jul 20, 2004 11:27 pm
Location: Seattle, WA
Contact:

Post by phaedrus »

davd_bob wrote:About the Celeron
The Mendocino core built on 25micron technology hits a wall between 550 and 625MHz. Pretty much any of them will run 550 nomater what speed they are "rated" for but almost none will go 625 mhz. If your system is running stable at 2.0volts try drop it to 1.85 and you may get a pleasant supprise.
Absolutely, every processor has this wall, and it comes from a number of effects. For instance, things like the finite amount of time that it takes for an individual transistor to switch (meaning, when the voltage on the gate of the FET is changed, it takes a finite amount of time for the body of the FET to change its conduction properties such that the transistor turns on or off). What this does is put a small delay in the signal for each transistor it passes through. Now, this is a tiny amount, but when you stack together enough transistors, the accumulated delay can get to be on the order of the clock frequency. When this happens, the processor will go out of sync with the clock and become unstable.

There are other things (lots of them, solid state and semiconductor physics gets pretty complex long before you hit the 25 micron level). I just put this forward as an example of a non-temperature limitation of a given processor.

There are other temperature dependancies that are not intuitive. The main problem with high temperature is electromigration, which doesn't effect the transistors themselves, but the interconnects (basically, the interconnects get hot enough that the current can push the metal around and destroy the connection, burning out the processor). The other interesting temperature effect is that as you decrease the temperature, the transistors' properties tend to improve greatly, and when you hit around liquid nitrogen temperature (77 K), the transistors become noticeably faster. It's quite amazing, actually. This effect is so strong, that research into superconducting computers (using Josephson Junctions) was abandoned simply because supercooled silicon worked nearly as well for much reduced cost (easier to manufacture, only needed to maintain LiN2 temperatures to get significant advantages and not LiHe (4.2 K), LiHe is much more expensive and much more difficult to handle (it diffuses through glass)).

The suggestion to drop the voltage to 1.85 is a good one, your power consumption will go down, and your stability likely won't suffer. Power, if you remember from electronics classes, is voltage times current. Drop the voltage, you drop the power consumption.

Jeff
"If it ain't broke, mod it till it is"
They said... and now my BP6 needs new processors... D'oh
Slackware Linux v10.1
Image
Post Reply